Additive Calibration Model for the Monitoring Data of PM2.5 and PM10 Based on ARIMA and Multiple Linear Regression

Abstract

Air pollution is harmful to the ecological environment and human health. PM2.5 and PM10 are particularly harmful to human health. Real-time monitoring of the concentration of PM2.5 and PM10 can grasp the air quality in time and take corresponding measures to the pollution sources. Monitoring data may be affected by meteorological factors, so we need to check and correct the monitoring data to improve its accuracy. The difference between the two groups was significant through exploratory analysis. The self-correlation analysis to the data of SDD showed it was high significant. So, ARIMA models were used by time series analysis to the data (A_i). Meteorological factors such as temperature were taken as independent variables, and the difference between the data of the two groups was taken as dependent variable. We established multivariate linear regression models (B_i). The additive calibration models were obtained (Y_i=A_i+B_i). The error analysis showed that the accuracies of PM2.5 and PM10 were improved, especially the calibration effect of PM10. Therefore, the additive calibration model based on ARIMA and multiple linear regression could effectively calibrate the monitoring data of SDD.